As international tax competition intensifies due to deepening economic integration, policy makers around the world find it increasingly important to know how the tax burden on labour and capital in their country compares to effective tax burdens in other countries. At the same time academic researchers have been looking for improved measures of effective tax rates in their efforts to document how international tax competition affects the level and structure of taxation, and for use in empirical analyses of the effects of taxes on economic behaviour and performance. As a result, the recent years have witnessed a proliferation of academic papers as well as national and international government reports offering new measures of effective tax rates on capital and labour. Because many of these measures are based on quite different methods, they often provide rather different estimates of the relative tax burden across countries. Moreover, the users of these estimates - in academic circles as well as policy circles - do not always seem to be aware of the different methodologies and assumptions underlying the different and often contradictory figures. Hence analysts and policy makers are far from having a clear picture of the level and evolution of effective tax rates in the OECD area. Against this background, this year's CESifo Venice Summer Institute Workshop in Public Economics focused on the measurement of effective tax rates. The purpose of the workshop was to provide an overview of the state of the art in this area and to offer and compare new and updated estimates of the tax burden on capital and labour and its evolution over time. An important goal was to clarify the methodological similarities and differences between the different effective tax rate measures which are currently in use. The workshop brought together leading academics in the field and economists who are engaged in producing effective tax rate measures for international organisations like the OECD and the EU. The following is a very brief summary of the papers presented at the workshop.
Michael P. Devereux opened the workshop by presenting two survey papers on the measurement of taxes on income from capital. His review explained the link between the neoclassical theory of investment and some of most well-known tax rate measures, stressing the distinction between the average effective corporate tax rate which is likely to govern decisions on the international location of investment, and the marginal effective corporate tax rate which will influence the volume of investment within a given country, once the location decision has been made. He also discussed how to account for personal taxes, alternative sources of finance, and risk. He then moved on to compare the 'forward-looking' effective tax rate measures based directly on economic theory with some widely used 'backward-looking' measures based on data on tax liabilities or revenues. In the final part of his presentation, he illustrated the properties of alternative measures of the taxation of income from capital by applying them to data for the UK. This exercise made clear how different approaches can give rise to very different quantitative estimates of tax rates.
As already hinted at by Michael Devereux, the treatment of firms' financing behaviour and their attitudes towards risk has been a weakness of the traditional approach to measuring effective tax rates on capital. In a joint paper with Giampaolo Arachi, Julian Alworth proposed a new method of measuring the effective tax rate on capital, based on the modern theory of corporate finance. The proposed measure explicitly accounts for the pricing of risk and implies that effective tax rates are uniquely linked to the company's optimal debt-equity ratio which is endogenously determined. A nice feature of the method is that it includes the traditional King-Fullerton effective tax rate measure as a special case. In a subsequent provocative paper Roger Gordon, Laura Kalambokidis and Joel Slemrod asked: "If capital income taxes are so high, why to we collect so little revenue?" Their point of departure was that the popular King-Fullerton method often yields rather high estimated marginal effective tax rates on capital, suggesting that taxes seriously distort the level and pattern of investment. Yet studies for several countries - including the authors' studies for the U.S. - have found that governments collect very little net revenue from capital income taxes. According to the authors this suggests that the King-Fullerton approach does not fully capture all of the key tax provisions, including the opportunities to shield income from tax through the use of debt finance and income shifting. The authors therefore proposed a new method intended to measure the difference between the actual tax revenue collected and the revenue which would be collected under a hypothetical R-base cash flow tax which is known to be non-distortionary and fall only on economic rents. In a detailed discussion, they argued that their new measure in most settings comes closer to the true marginal effective tax rate than some other widely used measures.
The following two papers presented the methodologies used by the OECD secretariat to estimate effective tax rates on capital and labour.
In a joint paper with Josette Rabesona, David Carey explained how the OECD Economics Department has extended the popular method of Mendoza, Razin and Tesar to improve the measurement of average effective tax rates on capital, labour and consumption, based on the OECD Revenue Statistics and the OECD National Income Accounts. The basic idea of the approach is to estimate that part of observed tax revenues which can be said to fall on capital, labour, and consumption, respectively, and to relate these revenue figures to the respective estimates of total capital income, total labour income, and total consumption to arrive at aggregate measures of average effective tax rates on these three main tax bases. David Carey discussed the many methological problems as well as the many data problems implied by this procedure for arriving at what is also called "implicit tax rates" or "tax ratios". He also presented new estimates suggesting that when a number of adjustments to the widely used tax rate measures of Mendoza, Razin and Tesar are made, the past quarter century does not seem to have been characterized by the fall in the relative tax burden on capital which emerges from the estimates of Mendoza et al.
The next paper, by Christopher Heady of the OECD Centre for Tax Policy and Administration, described an alternative method used by the OECD to measure the tax burden on labour, the so-called 'Taxing Wages' approach. The basic approach is to define a small number of 'typical families' in OECD member countries and to apply the tax rules for each countries to these family types in order to calculate the average and the marginal effective tax rates. To permit comparisons between those countries that mainly assist families through the tax system and those that mainly assist them with cash benefits, the calculations include universal family benefits paid in cash in respect of dependent children. The paper clarified a number of practical issues raised by the Taxing Wages approach, such as the calculation of gross wages, the choice of taxes and benefits to include, and the calculation of personal income taxes. It ended by presenting the latest OECD estimates of the direct tax burden on labour, including income tax plus employee and employer social security contributions.
The last paper on the first day of the workshop was written by Jakob de Haan, Jan-Egbert Sturm and Björn Volkerink and also dealt with the problem of measuring the tax burden on labour. The paper critically surveyed and discussed seven different macroeconomic indicators for the tax burden on labour which have been applied in the recent empirical literature. The authors found that the choice of the indicator for the tax burden on labour affects the conclusions from some recent empirical studies of the effects of taxation on employment and investment. However, they also showed that the correlation between most of the alternative measures of the average labour tax burden is quite high. Hence the particular choice of measure does not seem to be too important when one wants to study the trend in the evolution of the tax burden over time.
The second workshop day started with a presentation by Jim Hines who asked: what taxes are relevant when we try to measure the tax burden on international investment? In several previous papers, the author has provided ample evidence that the corporate income tax affects foreign investment and the amount of taxable profit reported. In line with this, existing measures of the tax burden on international investment focus almost entirely on the corporate income tax. But governments also tax local business activity with a number of other instruments such as property taxes, sales and value-added taxes, excise taxes, payroll taxes, etc. Hines developed a model to explore the circumstances under which these other taxes are likely to influence the pattern of international investment. His results indicated that taxes have the strongest effects when they burden one type of investor much more heavily than others.
The following paper by Harry Grubert likewise focused on the problems of measuring the tax burden on multinational companies. His rich paper illustrated how the actual effective tax rate on overall U.S. direct investment abroad is affected by income shifting through the relocation of assets, the relocation of debt, the share of foreign income repatriated as royalties, as well as several other factors. His empirical analysis particularly emphasised the importance of the provisions that govern the taxation of royalties, the use of tax haven finance subsidiaries, and the allocation of parent interest expenses to foreign income. Grubert also presented an empirical analysis of the way host governments tax different types of inward investment by U.S. multinationals. Among other things, he found evidence that host governments tend to favour more mobile activities and those activities that offer benefits to local factors such as labour. The overall impression left by Grubert's analysis was that producing a fully correct estimate of the effective tax burden on foreign direct investment is bound to be extremely difficult, given the complexity of the current systems of international taxation.
While a lot of scholarly effort has been invested in the development of measures of effective tax rates on physical capital, very little has been done to develop a method for estimating effective tax rates on human capital. In a joint paper with Jim Davies, Kirk Collins analysed a wide range of tax provisions which affect the incentive to invest in human capital. As a basis for the analysis, the paper developed a conceptual framework for measuring effective tax rates on the return to human capital investment. The proposed method distinguished between the direct effects of taxes and the investment incentives provided through the expenditure side of the public budget. Collins went on to illustrate how this framework can be put to work by applying it to Canadian data. The analysis revealed that, while the direct effective tax rate on the return to human capital are sizable although not as large as effective tax rates on physical capital, the net effective tax rate is on average negative when government education subsidies are accounted for.
The last two papers at the workshop illustrated the usefulness of micro data in the estimate of effective tax rates.
The first of these papers was given by Steven Clark of the OECD Centre for Tax Policy and Administration. He noted that measuring effective tax rates using actual tax revenues is attractive, since revenue figures capture the net effect of tax provisions and taxpayer behaviour that may be difficult to model. At the same time reliance on aggregate revenue and income data requires restrictive assumptions to separate the various income types and their respective tax burdens from each other. By contrast, micro-data may permit more direct measurement of the amount of tax revenue raised from various categories of income. The paper by Clark explained how micro-data can be used to assess average effective tax rates on labour, capital and transfer income, stressing the importance of matching taxpayer-level information to income flows. Clark's method also highlighted the importance of loss adjustments in measuring effective tax rates on capital income and gave evidence of significant variation in corporate effective tax rates by sector and firm asset size.
The issue of sector and size effects on effective corporate taxation was also the theme of the last workshop paper, presented by Gaëtan Nicodème from the European Commission. Using the BACH (Bank for the Accounts of Companies Harmonized) micro data base created by the EU Commission's Directorate General for Economic and Financial Affairs, he calculated the average effective corporate tax rate for each company as the ratio of taxes paid to the gross operating profit recorded in the harmonized accounts. Controlling for a number of other relevant factors, he then estimated sector and size effects on effective corporate tax rates for eleven European countries as well as the US and Japan. In line with Steven Clark's paper, the estimates by Nicodème revealed that there are large variations in effective tax rates across sectors and across companies of different size. While some of these differences may the intended result of targeted subsidies, others may reflect the unintended effects of an ill-designed discriminatory tax system. Effective tax rate measures based on micro data may thus help to identify tax non-neutralities which policy makers might want to remedy.
Overall, the workshop gave a good impression of the current status of theoretical and applied research on the measuring of tax burdens on capital and labour. CESifo intends to publish a book volume with the MIT Press based on the contributions to the workshop. It is hoped that this book will become a valuable hand-book for academics and policy makers seeking an overview of the state of the art of measuring effective tax rates and a clarification of the strengths and weaknesses of the many different methods currently used in this important area of applied economics.
RSS